Hello,
I have an NVME by Transcend that is in read only locked mode, not sure what it's story is - manufacturer states it is a protect function (controllers sensed NAND failing, blocks writes to preserve data but I think it is a controller firmware issue as it was barely used).

Was playing with it in unRAID test gig when I noticed that unAssigned devices happily attempts to wipe it and format it. It fails, but it doesn't really let me know it fails and the drive has a green dot.

This got me thinking, what happens if I had a raidz1 pool of 3x NVMEs and one of them does exactly this? Has anyone experienced this? Is RaidZ stable enough that I don't have to worry about loosing data? I know, back ups, back ups, but this would be for my cache pool, so backups aren't very practical in this example.

Currently running raidz mirror for cache pool for the above reasons but I would love to move to 3x raidz1 for some extra capacity.

Edited January 31Jan 31 by OrdinaryButt

If only a device fails, either completely or by going read-only, the pool would survive, and the device would need to be replaced.

7 hours ago, JorgeB said:

If only a device fails, either completely or by going read-only, the pool would survive, and the device would need to be replaced.

I know that is the expectation. However, this failed SSD behavior is weird. Unraid doesn't label it as bad, although it isn't part of a pool/array, so that might be way. But in addition to annasigned devices not recognizing it is bad, multiple linux and winndows tools (including gparted) pretended it is fine. Gparted even said it was secure wiped 100%.

16 hours ago, OrdinaryButt said:

Unraid doesn't label it as bad

There's currently no pool monitoring in Unraid, you can use a script to monitor them, last I heard this may change for 7.3

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-700582